The Use of Halide Charged Interleaves for Treatment of Iron Gall Ink

Home , Iron gall ink, Quill

The use of halide charged interleaves for treatment of iron gall ink damaged papers Véronique Rouchon, Maroussia Duranton, Oulfa Belhadj, Marthe Bastier-Deroches, Valéria Duplat, Charlotte Walbert, Birgit Hansen

To cite this version:

Véronique Rouchon, Maroussia Duranton, Oulfa Belhadj, Marthe Bastier-Deroches, Valéria Du- plat, et al.. The use of halide charged interleaves for treatment of iron gall ink damaged papers. Polymer Degradation and Stability, Elsevier, 2013, 98, pp.1339 - 1347. 10.1016/j.polymdegradstab.2013.03.028. hal-01435102

HAL Id: hal-01435102 https://hal.archives-ouvertes.fr/hal-01435102 Submitted on 13 Jan 2017

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights Author's personal copy

Polymer Degradation and Stability 98 (2013) 1339e1347

Contents lists available at SciVerse ScienceDirect

Polymer Degradation and Stability

journal homepage: www.elsevier.com/locate/polydegstab

The use of halide charged interleaves for treatment of iron gall ink damaged papers

Véronique Rouchon a,*, Maroussia Duranton a, Oulfa Belhadj a, Marthe Bastier-Deroches a, Valéria Duplat a,b, Charlotte Walbert a, Birgit Vinther Hansen b a Centre de Recherche sur la Conservation des Collections, Muséum national d’Histoire naturelle, Centre national de la recherche scientiﬁque, Ministère de la Culture et de la Communication, USR3224, CP21, 36 rue Geoffroy-Saint-Hilaire, 75005 Paris, France b The Royal Library, Preservation Dept., P.O. Box 2149, DK-1016 Copenhagen, Denmark article info abstract

Article history: In certain conditions, iron gall inks induce substantial damages on cellulosic supports. Existing curative Received 14 December 2012 treatments necessitate the immersion of the documents in aqueous solutions. This implies several un- Received in revised form desirable side effects: mechanical stress of the most damaged papers, substantial colour changes, and a 4 March 2013 drastic modification of the paper/ink composition. This work investigated an alternative treatment that Accepted 31 March 2013 consists in compressing the document between two interleaves that are charged with active compounds. Available online 17 April 2013 When this operation is operated in moist conditions, active compounds may migrate from the interleaves to the document, and thus achieve the treatment of the paper. As high humidity conditions may addi- Keywords: fi Halide tionally induce some migration of iron and acidic compounds out of the ink line, it rst appeared Iron gall ink necessary to research humidity conditions which minimize these undesirable migrations. This investi- Paper gation was performed on a set of 53 original manuscripts and led us to consider that a relative humidity Treatment of 80% was quite reasonable. The second part of the paper deals with the evaluation of interleaving Interleaf treatment efficiency. It was performed on laboratory samples, using several types of antioxidant halide Humidity salts (NaCl, NaBr, CaBr2) (halides are known to act as radical scavengers). These compounds were tested separately or in combination with calcium carbonate as an alkaline buffer. The samples were made of Whatman paper that was preliminary impregnated with iron gall ink. It has shown that a charge of 2% calcium carbonate in the interleaves has no impact on the treatment efficiency, which mainly depends upon the concentration of salts, the contact pressure and the duration of the treatment. Elemental measurements show that the mobility of the salts is much higher than that of iron, suggesting the possibility to find optimum treatment conditions that minimize the iron migration risk while achieving an effective migration of halides. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction acid), and gum arabic as a binder. They also contain large quantities of iron and present low pH values (in the range 1.5e3). As a result, 1.1. Iron gall inks: a major threat for western archives they can cause a major degradation of the cellulosic carrier [2,3]. This degradation results from the superimposition of two types In western countries, iron gall inks were used on a significant of phenomena: the first is mainly physical and corresponds to the part of archival patrimony dating from medieval times up to the migration of acidic and iron rich by-products out of the ink line [4]. beginning of the 20th century. These inks are made from a mixture It is, among others, provoked by exposure to high humidity con- [1] of iron salts (for example iron II sulphate heptahydrate), gall ditions [5]. The second is mainly chemical, and occurs in the paper nuts extracts (rich in polyphenolic carboxylic acids, such as gallic fibres that are in contact with the ink. This chemical degradation includes cellulose hydrolysis promoted by the acidity of the ink, and cellulose oxidation, catalysed by iron II and possibly imparted * Corresponding author. Tel.: þ33 1 4079 5303; fax: þ33 1 4079 5312. to hydroxyl radicals produced via Fenton mechanisms [6].Itwas E-mail addresses: [email protected] (V. Rouchon), [email protected] recently demonstrated that oxidation was the driving force of cel- (M. Duranton), [email protected] (O. Belhadj), [email protected] (M. Bastier-Deroches), [email protected] (V. Duplat), cha.walbert@ lulose depolymerization induced by iron gall inks [7]. This suggests gmail.com (C. Walbert), [email protected] (B.V. Hansen). that the oxidized cellulose reorganizes through chain scissions with

1340 V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347 a kinetic that is much faster than that of acidic hydrolysis. In 1.4. Aim of the study addition, it was shown that gallic acid was favouring oxidation phenomena [6], probably because of its high ability to reduce iron This study aims to investigate halide based interleaving treat- III into iron II [8,9]. ments using moderate conditions. Halide salts were chosen for this purpose and used separately or in combination with calcium car- 1.2. Available treatments and their side effects bonate. In addition to the aforementioned sodium bromide (NaBr), calcium bromide (CaBr2) was chosen because of the ability of cal- A significant amount of work has been realized in the last de- cium to bind carboxylic groups on cellulose chains [31], which may cades to find chemical treatments capable of limiting iron gall ink improve the treatment efficacy. Sodium chloride (NaCl) was addi- damages. The calcium phytate process proposed in the 90’s [10] has tionally included because chloride, already highlighted as an effi- been the most widely tested [11]. Although its efficiency is well cient anti-oxidant [21,24], is smaller than bromide, and thus may established, this treatment is rarely used, mainly because it re- migrate more easily. quires immersions in aqueous solutions, which provokes drastic When manuscripts are exposed to high humidity conditions, modifications of the paper’s appearance, of its chemical composi- brown ink compounds may additionally migrate out of the ink line tion [12e14] and mechanical damages for brittle papers [15]. These and through the paper. These unaesthetic brown halos are often side effects are often found unacceptable. accompanied by a substantial migration of iron [5] and should For several decades, halides are known to decompose hydrogen therefore be considered as potentially detrimental to the conser- peroxide [16]. This property is probably correlated to their ability vation of the carrier. We thus dedicated the first part of this work to to react with hydroxyl radicals thus acting as radical scavengers the risk assessment of these migration phenomena. [17e19]. In the 2000’s, a major effort was paid for the testing of The second part of this research deals with the monitoring of the non-aqueous treatments using halide products as anti-oxidants. treatment. Several types of interleaves and several types of treat- Several studies report the efficiency of bromide and chloride ment were implemented in order to identify the most effective compounds prepared in several organic solvents [20e25]. These ones. results are encouraging. However, in all these studies, halide based treatments were implemented after a conventional aqueous 2. Experimental deacidification, meaning that the global treatment remains water based with all relating side effects. 2.1. Methodology to evaluate the risk of ink migration related to high humidity exposure 1.3. Interleaving techniques The risk of ink migration was evaluated on a set of 53 original Interleaving techniques consist in placing the document in manuscripts dating from 1818 to 1896, and originating from the contact with an interleaving paper impregnated with some active south of France. These manuscripts were formally chosen to study compounds. The ensemble is then pressed and kept at high relative the migration risk of aqueous treatments and are therefore humidity (RH) for several days. This high level of humidity pro- described in detail elsewhere [5]. Several pieces of approximately vokes the migration of active compounds from the interleaves to 2e4cm2 were sampled on each manuscript. Writing with a quill the document, thus achieving the paper treatment. It can addi- pen produces an uneven ink line and the line width may consid- tionally provoke the migration of acidic and/or iron rich by- erably vary along the writing according to the style of the writer. It products from the document to the interleaves [26]. was observed that thick lines were more prone to migration, Interleaving techniques were first implemented on printed probably because they contain a higher quantity of ink. Attention documents for de-acidification purposes. Their capacity to increase was therefore paid to sample pieces that do not only contain thin the pH of acidic papers has been demonstrated [26] in the context lines, but also thick lines. Visible migrations were evaluated with of a patent deposited in 1993. The efficiency of the treatment can macro-pictures taken with a digital camera. It became obvious that be improved by the presence of salts and drastically depends on a conventional light booth was too large compared to the di- the pressure and RH that is applied during treatment. Best results mensions of our samples to allow the capture of reproducible are achieved with a humidity of 97% RH and a pressure above macro-pictures. We thus designed a specific three compartment 70 kPa (this value corresponds to a mass of 420 kg placed on an A4 light booth with a digital camera placed on the top (Fig. 1). A 4300 K format sheet). The possibility to use less drastic conditions was diffuse illumination was obtained using four tungsten halogen researched by Hanus et al. [27] who report that a pH increase of 1e 2 points is achievable at 92% RH using interleaves that contain a high alkaline reserve (approx. 15% of CaCO3) and applying a pressure of 7 kPa. All these conditions are not realistic for original manuscripts, because (i) there is a substantial risk of ink migration if the paper is exposed to such a high humidity [5], and (ii) such a high pressure will certainly flatten the paper surface, which is detrimental to the document in the case of laid, watermarked, or dry-stamped papers. Additionally, the treatment of iron gall ink manuscripts necessi- tates the use of an anti-oxidant. For this purpose, halide compounds appeared promising: Interleaves charged with sodium bromide and calcium carbonate appeared efficient in limiting the paper decay provoked by iron gall inks [28,29]. More recently, the use of tetrabutylammonium bromide enabled to delay paper oxidation in a mildly alkaline environment [30] (which is unfor- tunately not the case of untreated original manuscripts that are Fig. 1. Schema of the light booth. (1) Sample; (2) white walls; (3) ground glasses; (4) acidic). black walls; (5) tungsten halogen lights; (6) digital camera. Author's personal copy

V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347 1341 lights that illuminate through a ground glass window a central stabilization of the chamber was difficult to monitor. For this white walled box. To absorb the backward emitted yellow light the reason, the experiments were duplicated by the use of small en- walls of the two light compartments were painted black. The closures with saturated salt solutions [32] (Table 1). This type of reproducibility of the lighting was checked with a 1 2 cm colour installation was more accurate and easier to manage. Throughout test chart placed next to the sample. these experiments, temperature and RH were monitored with The visible migrations of ink were identified visually by three Hygro Buttons (Plug and Track, Proges Plus, precision: tempera- paper conservators who have a professional expertise in paper ture 1 C, RH 2%). examination. They are trained in distinguishing small colour vari- ations on virgin paper (1 < DE < 3). The three agreed to the final 2.2. Preparation of inked samples for interleaving treatments evaluation. Visible migrations were classified in two categories: “transversal Samples consisted of paper sheets impregnated with a diluted migrations” concerned the transfer of coloured compounds iron gall ink prepared with pure laboratory products: gallic acid through the paper (verso side), whereas “lateral migrations” dealt monohydrate (Aldrich, 398225), 0.9 g L 1 and Fe(II) sulphate hep- more specifically with the formation of halos around the ink line tahydrate (Aldrich, 215422), 4 g L 1. The choice of ingredients and (recto side). For each category, it was necessary to define two levels their concentrations are already explained elsewhere [7]. The ink of observation (Fig. 2): migrations were considered to be “obvious” was stirred in a closed vessel for 3 days until it reached a pH of when they were quickly noticeable by naked eye and “subtle” when 3.0 0.1. The paper sheets of Whatman no 1 (6 9cm2)were a careful examination of macro-pictures was necessary to attest the immersed for 10 min in the ink solution, respecting a maximum changes. ratio of 4 cm2 of paper per 1 mL of ink. After immersion, the excess Investigated humidity conditions ranged from 80% to 97% RH. At of solution was mopped up in a similar way to that depicted in the this level of humidity, there is a significant risk of mould growth if Cobb test [33]: the paper was placed between two blotting papers the duration of the treatment is too long. The maximum tolerable and the assembled sheets were squeezed by the use of a 10 kg roll. duration above which the mould risk becomes significant was This procedure enables to achieve an even distribution of the ink. chosen (Table 1). These values were estimated by the use of the The paper was then dried in ambient conditions. Dew point calculator software (Image Permanence Institute, Rochester, USA). In a first step of the evaluation, the samples were simply 2.3. Preparation of the interleaves placed on a non-woven polyester film and in a climatic chamber (VC 0020, Vötsch). However, as the apparatus did not cover hu- The interleaves were made of Whatman no 1 (13 13 cm2) midity conditions over 85% at 20 C, it appeared necessary to impregnated with different types of solutions as depicted in add water containers in the chamber in order to reach the Table 2. Each impregnation was performed using 100 mL of solution highest humidity conditions. This method was laborious and the per interleaf and each immersion lasted for 5 min.

Fig. 2. The different levels of lateral and transversal migration of brown products. (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of this article.) Author's personal copy

1342 V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347

Table 1 Number of original samples on which brown migrations are visible. Total number of original manuscripts: 53.

Relative humidity Duration Type of migration Small enclosures with saturated solutions Climatic chamber

Type of salt Lateral migrations Transversal migrations Lateral migrations Transversal migrations

97% 2 days No K2SO4 48 18 50 22 Subtle 5 30 3 28 Obvious 0 5 0 3

93% 4 days No KNO3 53 39 53 45 Subtle 0 14 0 8 Obvious 0 0 0 0 85% 6 days No KCl 53 53 53 53 Subtle 0 0 0 0 Obvious 0 0 0 0

80% 14 days No NH4SO4 53 53 53 53 Subtle 0 0 0 0 Obvious 0 0 0 0

Calcium carbonate (CaCO3) was introduced in the paper using presence would prevent fluctuation of humidity if the treatment saturated solutions of calcium hydroxide (Ca(OH)2). During drying, was not performed in a laboratory, but in a conservation workshop. calcium hydroxide reacts with carbon dioxide, leading to the for- The assembly was introduced into a polyester folder and pressed mation of calcium carbonate. However, the low solubility of cal- between two glass plates with a weight positioned on the top. cium hydroxide (2 g L 1) makes impossible to deposit in one Finally, the stack was placed in a climatic chamber (V0020, Vötsch) immersion the desired amount of calcium carbonate (>2% w/w). at the desired RH for several days to assure constant humidity Impregnations with calcium hydroxide were then duplicated and conditions during the contact. During the treatments, the temper- the paper was dried in between in order to provoke the formation ature and RH were controlled with Hygro Buttons (Plug and Track, of calcium carbonate. Proges Plus). Solutions of halide salts were prepared with the same molar concentration of chloride or bromide (1 M). After immersion, the 2.5. Artificial ageing and measurement of paper decay sheets were placed between two Cobb blotting papers and the excess of solution was mopped up as depicted above for the Before treatment, inked papers were aged in a climatic chamber preparation of inked samples. (V0020, Vötsch). This causes an initial degradation of paper in order to resemble damaged manuscripts. This ageing should however 2.4. Implementation of the treatments and preliminary testing remain moderate, because the samples must be re-aged after treatment to assess their evolution over time and the effectiveness All treatments were performed at room temperature 23 C. of the treatment. To this respect, standard [34] moist heat ageing When the papers are stacked and pressed, the time required to conditions (80 C, 65% RH) were found too drastic, since the inked reach equilibrium with ambient humidity can be extremely long. In paper was becoming completely brittle in less than one day. Milder some preliminary experiments, it was found that the relative hu- ageing conditions were instead used (70 C, 65% RH) and a loss of midity within a stack of paper that is pressed for several days could approx. 15% of mechanical properties was achieved within 3 days. remain about 10% below the ambient humidity. To avoid this After treatment, the inked papers were separated from the in- discrepancy, all the material and equipment used for the contact terleaves and aged again in the same conditions for approximately wereis pre-conditioned for at least 2 days to obtain the desired 2 months. humidity, without any pressure being applied. This step ensures As conservation treatments aim to limit physical decay of the that the samples are already in equilibrium with the desired paper, mechanical testing appeared meaningful for the evaluation moisture content before the treatment. of treatment efficiency. The decay of paper versus ageing was The contact between the samples and the interleaves was evaluated with a Zero-span tensile tester [35] (Pulmac, TS-100). The established according to the schema depicted in Fig. 3. The sample test was performed on dried papers, pre-conditioned for at least 4 h was sandwiched between two interleaves and two Cobb blotting at 23 C and 50% RH. It consists in measuring the failure load papers. Cobb blotting papers were not used in previous studies [29], but were introduced in this work. It was expected that their

Table 2 List of interleaves and solution used. All halide solutions were prepared at 1 M concentration of halide.

Name First bath Second bath (if any) Third bath (if any)

Salt (molar conc.) Roll Salt (molar Roll Salt (molar Roll conc.) conc.)

Ca(OH)2 CaOH2 (0.027) No CaOH2 (0.027) No ee NaBr NaBr (1.0) Yes eeee NaCl NaCl (1.0) Yes eeee

CaBr2 CaBr2 (0.5) Yes eeee Fig. 3. Schema of interleaving treatments. The document (1), the two charged in- Ca(OH)2/NaBr CaOH2 (0.027) No CaOH2 (0.027) No NaBr (1.0) Yes terleaves (2), and the two blotting papers (3) are preconditioned at 80% RH, then Ca(OH)2/NaCl CaOH2 (0.027) No CaOH2 (0.027) No NaCl (1.0) Yes placed in a plastic folder (4) and pressed between two glass plates (5) with a load (6) Ca(OH)2/CaBr2 CaOH2 (0.027) No CaOH2 (0.027) No CaBr2 (0.5) Yes placed on the top. Author's personal copy

V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347 1343 necessary to break a strip of paper maintained by adjacent jaws. and climatic chamber), despite some small differences that are The load is expressed in kg per 15 mm of strip width. For each probably attributable to the heterogeneity of the manuscripts. sample, 10 measurements were performed and the average was When exposed at humidity over 90% RH, a significant number of calculated. Zero-span testing presents several advantages with documents are subjected to migration phenomena. On these sam- respect to other mechanical testing: (a) it requires a small quantity ples, the condensation of water in the micro(nano)-porous struc- of material only; (b) on Whatman no 1 paper, this test shows a ture of paper is probably sufficient to provoke the dissolution and better correlation with cellulose molecular mass Mw and less the migration of ink by-products [45]. On the contrary, the 53 dispersive measurements than tensile strength testing [36]. original samples could be exposed to 80% RH for 14 days or to 85% Mechanical tests were completed with pH measurements per- RH for 6 days without any visible migration on the recto and on the formed on cold extracts prepared at ambient temperature with verso of the ink lines. 0.25 g of paper in 12.5 mL of decarbonated ultrapure water [37]. The Elemental mappings were performed on a selection of samples. pH was measured 1 h after the immersion of the papers. On all samples showing obvious visible migrations, the distribution of iron was systematically superimposing that of brown halos, 2.6. Elemental measurement showing that iron migrates with brown products. The samples on which no elemental migration of iron was noticeable corresponded In order to evaluate elemental migration occurring on original to those showing no visible change or the most subtle of the visible manuscripts after exposure to high humidity conditions, a scanning migrations. On these samples, the iron migration (if any) was too electron microscope coupled with a dispersive X-ray detection low to be highlighted by SELM/EDS. spectrometer (SEM/EDS, JEOL JSM-5410LV, Oxford Instrument Link These observations are consistent with a previous work dealing Pentafet) was used. Specific attention was paid to the elements that with the evaluation of aqueous treatment side effects [5] which are originally contained in the ink (mainly iron, sulphur, potassium, showed that iron migrations, when detected by SEM/EDS, were and calcium). Experimental conditions used for elemental mapping systematically superimposing that of brown products. were the following: 30 keV beam, diaphragm 2, low vacuum 20 Pa, In order to estimate if a low amount of iron may migrate at 80% no sample preparation, acquisition time between one and 3 h. RH, we opted for a more sensitive (although not quantitative) References of iron sulphate impregnated papers (whose iron con- methodology using Bathophenanthroline impregnated papers. On tent was preliminary determined by atomic absorption spectrom- these samples, it was confirmed (Fig. 4) that there is a significant etry [38]) were used to estimate the low detection limit of SEM/ risk of iron migration at 97% RH for 1 day, whereas no iron EDS. It was close to 2 mg of iron per gram of paper, a value signif- migration could be observed at 80% RH for 10 days. icantly superior to the residual average amount of iron deposited in It could be objected that a great variety of iron gall inks exist and the paper during its manufacture (estimated to be in the range of some may exhibit a different behaviour from that of our reduced 0.1e0.4 mg/g on well preserved papers [39]). selection. It is however not realistic to test all kinds of original The possible migration of a low amount of iron (below 2 mg g 1) manuscripts before designing a conservation treatment. Our ob- was further investigated with a colour test, using the property of servations are consistent and led us to consider that the risk of ink Bathophenanthroline (a water insoluble indicator) of forming a migration remains relatively low during an exposure at 80% RH for purple complex/precipitate with iron II ions [40,41]. Bath- 10 days. ophenanthroline impregnated papers (solution used for the impregnation: 1.6 g L 1 in 99.9% ethanol) are largely used in the 3.2. Control of the initial composition of the interleaves field of paper conservation for the identification of iron gall inks [42,43] and for the monitoring of the phytate treatment [11]. In this In newly prepared interleaves, elemental concentrations of ha- study, we used Bathophenanthroline impregnated papers that are lides (Table 3) are close to 1000 200 mmole g 1 and quite similar stamped with an iron gall ink. These samples, which detailed from one salt to another. Moreover, the stoichiometric ratios be- preparation can be found elsewhere [44], were specifically tween the elements deposited in the interleaves remain close to designed to identify the risk of iron migration when aqueous ad- those of impregnating solutions. This observation is accounted to hesives are applied on manuscripts. They are capable to highlight the fact that Whatman paper is both a porous and absorbent ma- the presence of very small concentrations of iron II, close to terial that retains a certain amount of solution when removed from 0.06 mg g 1 [11]. the baths. This amount is estimated to be 1.15 mL of solution per Finally, in order to determine the average elemental contents of gram of paper when a Cobb roll is used for the removal of the inked papers and interleaves, the papers were mineralized in surplus of solution (impregnation with NaBr, CaBr2, and NaCl). This concentrated acid. Elemental concentration in iron, calcium, so- amount was used to estimate a concentration of halide of dium, bromide, and chloride was measured by ICP-AES (ICAP 6300, 1150 mmole g 1, a value close to the average aforementioned Thermofisher Scientific). Elemental concentration in sulphur was measurements (1000 200 mmole g 1). It is also reasonable to evaluated with a UV fluorescence nitrogen analyser (TN/TS 3000, consider that interleaves elemental contents correspond to the Thermofisher Scientific), after the total combustion of approx. 5 mg deposits brought by the solution that remains in the paper after of samples at 1000 C in the presence of argon and oxygen (in order immersion. to assure the conversion of sulphur into sulphur dioxide). All This observation is not valid when considering calcium con- samples were triplicated and the average was considered. tents. After impregnation with Ca(OH)2 solutions, the surplus of water is not removed and the remaining solution (2.7 mL per gram 3. Results of paper) should theoretically deposit 73 mmole g 1 of calcium. When considering, on respectively “Ca(OH)2”, “Ca(OH)2/NaBr” and 3.1. Evaluation of visible side effects induced by high humidity “Ca(OH)2/NaCl” interleaves, that the deposits of the two successive exposure Ca(OH)2 solutions are superimposing, we may estimate the final calcium content at 146 mmole g 1, a value inferior to our mea- The methodology used for the evaluation of migration of brown surements (respectively 270, 220 and 240 mmole g 1). This point is products appears satisfactorily reproducible as the results of the consistent with previous observations formulated on papers two sets of measurements are similar (see Table 1: small enclosures impregnated with other types of calcium containing solutions [11] Author's personal copy

1344 V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347

Fig. 4. Migrations of iron II detected by the use of the Bathophenanthroline test paper. These papers were stamped with iron gall ink then cut into several pieces. The central parts were exposed to high humidity conditions, whereas the upper and lower parts were not treated. (a): 97% RH, 1 day, recto side; (b): 97% RH, 1 day, verso side; (c): 80% RH, 10 days, recto side; (d): 80%, 10 days, verso side. and highlights a strong interaction between calcium and cellulose 3.4. Evaluation of the treatment efficiency during immersion. The interleaves depicted above were firstly used to implement 3.3. Preliminary testing of efficiency the treatment at 80% RH, with a pressure of 4 kPa, and testing respectively a short exposure period (2 days) and a longer exposure Following the results of previous work [29], some primary period (10 days). The duration of the exposure appears a determi- fi treatments were performed during a 10 day period at 80% RH, using nant parameter: 2 day exposures were inef cient in delaying the a global pressure of 2.5 kPa (15 kg on an A4 format sheet). In these loss of mechanical properties (Fig. 5a), and in increasing the pH of the paper which remains close to 4.0 0.3 (not shown). On the conditions, the “Ca(OH)2/NaBr” interleaves were inefficient in delaying the paper decay, probably because (a) the interleaves contrary, when the duration of exposure is extended to 10 days prepared in Vinther Hansen’s work were more concentrated than (Fig. 5b), some treatments were limiting the loss of mechanical this series (they were not pressed with a Cobb roll), and (b) the properties. The best results were achieved with sodium chloride “ ” “ ” addition of Cobb blotting papers during the contact enables salts to impregnated interleaves ( NaCl and Ca(OH)2/NaCl ) whereas migrate in an additional medium and thus less in the inked paper. calcium carbonate and calcium bromide impregnated interleaves “ ” “ ” “ ” The treatment became efficient when repeated at a higher hu- ( Ca(OH)2 , CaBr2 and Ca(OH)2/CaBr2 ) seemed to have no effect midity (95% RH). However, we meanwhile found that such a high at all. humidity was provoking a substantial migration of iron and col- Table 4 shows pH measurements performed on inked paper oured compounds out of the ink line. This point was clearly after treatment and at the end of artificial ageing. First, the pH of prohibitive. It was therefore decided to remain at 80% RH and to investigate higher pressures. However, as the flattening of laid, watermarked or dry-stamped papers was undesirable, high pressure values such as those used for de-acidification treatments [26,27] were not chosen. It was found preferable to opt for an intermediate pressure of 4 kPa (25 kg on an A4 format sheet).

Table 3 Elemental composition of the interleaves before treatments. Elemental concentrations are measured with a precision of 10%. “Whatman” refers to virgin Whatman no 1 paper. Interleaves were prepared as depicted in Table 2.

Name Elemental concentration (micromole per gram of paper)

Na Ca Br Cl

Whatman 2 1 <0.5 <0.1

Ca(OH)2 e 270 ee NaBr 740 e 860 e NaCl 1040 ee 980 Fig. 5. Mechanical decay versus time of samples treated at 80% RH. All samples were CaBr2 e 550 1050 e ﬁrst maintained at 23 C, 80% RH during the treatments then aged artiﬁcially at 70 C, Ca(OH)2/NaBr 980 220 1160 e 65% RH. All treatments were performed using a pressure of 4 kPa for a period of 2 days Ca(OH)2/NaCl 1160 240 e 1200 (a) or 10 days (b). The graph plots the average of ten measurements. The error bars Ca(OH)2/CaBr2 e 770 830 e correspond to the standard deviation of these measurements. Author's personal copy

V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347 1345

Table 4 Table 5 pH and elemental composition of the inked samples after treatment. The treatments Iron and sulphur concentration of inked samples and interleaves after treatment. were performed for 10 days, with a pressure of 4 kPa, and using the interleaves Elemental concentrations are measured with a precision of 20%. depicted in Table 3. Precision of pH measurements 0.5; precision of elemental measurements 15%. The pH was measured after treatment and after 43 days of Micromole per gram artiﬁcial ageing (70 C, 65% RH). “No salt” refers to reference inked papers that of paper where treated with virgin Whatman papers. Fe S

Type of interleaf pH Concentration (micromole Inked samples e 16 15

used for the per gram of paper) Interleaves (after Ca(OH)2 0.14 0.9 treatment treatment) NaBr 0.30 1.6 After After ageing Na Ca Br Cl NaCl 0.09 1.1 treatment CaBr2 0.09 0.4 No salt 3.8 4.1 2 1 <0.5 <0.1 Ca(OH)2/NaBr 0.20 2.0 CaOH2 3.9 4.0 e 2.9 ee Ca(OH)2/NaCl 0.13 1.0 NaBr 4.3 5.3 290 e 200 e Ca(OH)2/CaBr2 0.13 0.6 NaCl 4.9 5.0 270 ee 420

CaBr2 3.8 4.7 e 47 150 e Ca(OH)2/NaBr 4.3 5.6 300 2.5 280 e e Ca(OH)2/NaCl 5.9 5.1 250 2.5 520 This is coherent with pH and elemental measurements (Table 4) e e Ca(OH)2/CaBr2 3.8 5.0 75 310 which show, as discussed above, that calcium does not migrate easily to the inked paper. For this reason, the “Ca(OH)2/CaBr2”, “Ca(OH)2/NaBr”, and “Ca(OH)2/NaCl” interleaves will be removed untreated samples (3.8) remained constant with ageing, contrarily from the rest of our work. to the commonly acknowledged idea that acidic papers grow more Considering that a large proportion of halide migrates from the “ ” acidic as they age [46]. Second, Ca(OH)2 interleaves appeared interleaves to the inked paper, one can a priori not exclude that fi absolutely inef cient in raising up the pH of inked papers or some proportion of iron and sulphur migrates from the inked paper delaying its degradation. More generally, the pH increase of inked to the interleaves. The concentrations of iron measured in the in- papers is not related to the presence of calcium carbonate in interleaves after treatment, reported in Table 5, show that approx. terleaves and the addition of calcium carbonate does not appear to 0.6%e2% of the initial amount of iron present in the inked samples have any positive impact. migrates to each interleaf, meaning that the iron that is present in “ ” For halide treatments that seem to have no effect ( CaBr2 and the inked samples is much less mobile than halides. This point is “ ” Ca(OH)2/CaBr2 ), the pH after treatment remains close to its initial crucial, because it shows that it should be possible to find experi- value (3.8), but rises up during ageing. For the most effective halide mental conditions that minimize the risk of original inks migration “ ” “ ” fi treatments ( NaCl and Ca(OH)2/NaCl ), the pH signi cantly in- while allowing halides migration. e creases during treatment up to a range of 4.9 5.9 and remains It can be seen in Table 5 that sulphur migrates more easily than fi stable through ageing. Thus the treatment ef ciency seems, to a iron: 3%e13% of the initial amount of sulphur present in inked certain extent, to be correlated to an increase of pH during treat- samples migrates to each interleaf. This point corroborates ment. Should this pH increase be considered as the major expla- elemental measurements previously performed on laboratory [47] fi nation of the treatment ef ciency? Is it simply a side effect? and original samples [48], that systematically pointed out a sub- fi Surprisingly, after 43 days of arti cial ageing, all papers treated sequent migration of sulphate ions out of the ink lines. with halide containing interleaves present the similar pH values (5.1 0.5) (Table 4) regardless to the efficiency of the treatment. The comparison of elemental concentration of inked samples after treatment (Table 4) with that of interleaves before treatment (Table 3) shows that the migration of halides is substantial: for instance, the average amount of halide migrating to inked samples during the treatment with “Ca(OH)2/NaCl” interleaves (520 mmole g 1) represents approximately 40% of the initial concentration of one interleaf (see Table 3, 1200 mmole g 1). Consid- ering that the samples were sandwiched between two interleaves, we can deduce that approx. 20% of chlorine present in each interleaf migrates to the inked paper, thus illustrating the mobility of chloride ions. The mobility of bromine and sodium is slightly inferior but remains important, as approx. 10%e15% of the initial amount of bromine and sodium present in each interleaf migrates to inked samples. On the contrary, the mobility of calcium is much lower: considering the interleaves “Ca(OH)2”, which are supposed to contain calcium carbonate, less than 1% of the initial amount of calcium migrates into the samples. This ratio increases up to approx. 5% in the case of “Ca(OH)2/CaBr2” and “CaBr2” interleaves, probably because calcium bromide dissociates more easily than calcium carbonate. However, we can notice that inked papers treated with “Ca(OH)2/CaBr2” and “CaBr2” interleaves contain, after treatment, between 3 and 4 times less calcium than bromine, Fig. 6. Influence of the concentration of salts in the interleaves. All samples were first meaning that calcium, when liberated by the dissociation of cal- maintained at 23 C, 80% RH during the treatments then aged artificially at 70 C, 65% RH. All treatments were performed using a pressure of 4 kPa. Concentration of salt cium bromide, are less inclined for migration than bromine. used for the impregnation of the interleaves: 0.2 M (a) and 5 M (b). The graph plots the On Fig. 5 it can be seen that the addition of calcium carbonate in average of ten measurements. The error bars correspond to the standard deviation of the interleaves does not affect the behaviour of treated samples. these measurements. Author's personal copy

1346 V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347

Fig. 7. Aspect of the paper surface under grazing light. Left: virgin Whatman paper; right: interleaf prepared with a 5 M NaCl solution.

3.5. Inﬂuence of the concentration of halide in the interleaves 4. Conclusions

In order to evaluate if the efficiency of the treatment could be This work was undertaken to evaluate the possibility of imple- improved by the use of more concentrated interleaves or if it re- menting interleaving techniques for the treatment of papers mains efficient with the use of more diluted interleaves, two new damaged by iron gall inks. We first evaluated the risk of ink sets of interleaves were prepared, in the same way as depicted migration at high humidity conditions on a set of 53 original above, yet using 5 times more concentrated (diluted) solutions. The manuscripts and found that an exposure at 80% RH for 10 days was treatments were implemented with the same conditions as above satisfactorily minimizing this risk. Treatments were thus imple- (80% RH, 10 days, 4 kPa). mented at this humidity level. The results obtained with these new sets of interleaves are It was shown that halides easily migrate from the interleaves to summarized in Fig. 6. Diluted interleaves appeared totally ineffi- the treated paper and that they are capable to delay the paper cient in delaying the paper decay (Fig. 6a). On the contrary, the decay, even in middle acidic environment. This stabilizing effect use of more concentrated interleaves significantly improved the might be correlated to an increase of pH during the treatment, but efficiency of the treatment (Fig. 6b). “CaBr2” and “NaBr” in- is not accounted to the calcium carbonate contained in the interleaves give better results: after 60 days of artificial ageing, the terleaves. Indeed, the presence of calcium carbonate in the in- zero-span breaking load of treated samples has decreased by terleaves did not have any measurable effect on the treatment approx. 30% while it reaches its minimum value on untreated efficiency, probably because calcium carbonate does not dissolve samples. easily during the treatment. Independently of the salt that is considered, the treatments with In this work, several factors were tested for a better monitoring concentrated interleaves (5 M solutions) provoked on all inked of the treatment: firstly the treatment should be sufficiently long samples a substantial pH increase, up to values ranging from 6.0 to and cannot be reduced from 10 days to 2 days. Secondly, the con- 6.5. Contrary to our previous observations (see Section 3.4 and centration of salts in the interleaves should be sufficiently high: Table 4), the pH of treated papers decreased with ageing and using 0.2 M solutions for the preparation of interleaves makes the reached pH values ranging 5.0 to 5.5 after 60 days of artificial treatment inefficient; using 1 M solutions makes the treatment ageing. Surprisingly, these values are similar to those presented in partially efficient for some salts; using 5 M solutions makes the Table 4 (pH after ageing). treatment efficient for all salts but leads to substantial undesirable Despite the fact that concentrated interleaves prepared with side effects on some samples. These considerations show that the 5 M solutions make the treatment more efficient, the work was concentrations of halides which migrate to the paper during the not pursued with such high concentrations, and this for two treatments is a key factor for the monitoring of the treatment reasons: first, in the case of NaCl, the high concentration of salt efficiency. provoked some mechanical tensions probably attributable to Encouragingly, we noticed that the migration of iron from the crystalline formation. The paper surface became uneven (Fig. 7). samples to the interleaves was largely less than that of halide This point jeopardized the treatment efficacy as it was not compounds from the interleaves to the sample. Experimental possible to achieve a satisfactory contact (see Fig. 6 “NaCl” conditions should be found that minimize the migration risk of iron treatment). Second, it can be observed in Fig. 6 that the treatment while enabling that of halides. modifies the initial mechanical properties of the paper: the initial It appears now necessary to pursue the effort by examining the zero-span breaking load is close to 8.2 0.5 on reference effect of the treatment on original valueless manuscripts. Their Whatman paper, and decreases slightly down to 7.4 0.5 on the composition is more complex than that of our laboratory samples, paper that are treated by “NaBr” and “CaBr2” interleaves. This which may raise new difficulties: the paper size for instance may modification of mechanical properties, which probably results drastically change migration phenomena. Finally, side effects, such from a large migration of salts, is not researched on original as the formation of salt efflorescence, or the flattening of laid, manuscripts. watermarked, or dry-stamped papers should be evaluated on Author's personal copy

V. Rouchon et al. / Polymer Degradation and Stability 98 (2013) 1339e1347 1347 original valueless manuscripts before implementing the treatment [21] Malesic J, Kolar J, Strlic M, Polanc S. The use of halides for stabilisation of e on valuable collections. iron gall ink containing paper the pronounced effect of cation. e-PS 2005; 2:13e8. [22] Malesic J, Kolar J, Strlic M, Polanc S. The influence of halide and pseudo-halide antioxidants in Fenton-like reaction systems. Acta Chim Slov 2006;53:450e6. Acknowledgement [23] Kolar J, Strlic M, Budnar M, Malesic J, Selih VS, Simcic J. Stabilisation of cor- rosive iron gall inks. Acta Chim Slov 2003;50:763e70. This study was funded by the French National Program of [24] Kolar J, Mozir A, Balazic A, Strlic M, Ceres G, Conte V, et al. New antioxidants for the treatment of transition metal containing inks and pigments. Restau- Research for Cultural Heritage (PNRCC 2010-2011). rator 2008;29:184e98. [25] Ceres G, Conte V, Mirruzzo V, Kolar J, Strlic M. Imidazolium-based ionic liquids for the efficient treatment of iron gall inked papers. ChemSusChem 2008;1:921e6. References [26] Page DH, Scallan MA, Middelton SR, Zou X. Method for the deacidification of papers and books. In: Patent US, editor. USA: Pulp and Paper Research Insti- [1] Stijnman A. Iron gall inks in history: ingredients and production. In: tute of Canada; 1995. Kolar J, Strlic M, editors. Iron gall inks: on manufacture, characterisation, [27] Hanus J, Minarikova J. Deacidification without equipment and money e degradation and stabilisation. Ljubljana: National and University Library; dream or reality? In: Conservation ICf, editor. 13th Triennial meeting, ICOM- 2006. p. 25e68. CC. Rio de Janeiro: James and James Ltd.; 2002. p. 603e8. [2] Kolar J, Strlic M, editors. Ageing and stabilisation of paper. Ljubljana: National [28] Hansen BV. Study of interleaving as a mean to reduce ink corrosion in man- and University Library; 2005. uscripts. In: Kolar J, Strilic M, Wouters I, editors. ICCOM-CC graphic document [3] Banik G, Weber H, editors. Tintenfrassschäden und ihre Behandlung. Stutt- meeting. Ljubljana: National and University Library; 2004. p. 30e1. gart: Kohlhammer; 1999. [29] Hansen BV. Improving ageing properties of paper with iron gall ink by [4] Remazeilles C, Rouchon Quillet V, Bernard J, Calligaro T, Dran JC, Salomon J, interleaving with papers impregnated with alkaline buffer and antioxidant. et al. Influence of gum arabic on iron gall ink corrosion, part II: observation Restaurator 2005;26:190e202. and elemental analysis of originals. Restaurator 2005;26:118e33. [30] Malesic J, Kocar D, Fabjan AB. Stabilization of copper- and iron-containing [5] Rouchon V, Durocher B, Pellizzi E, Stordiau-Pallot J. The water sensitivity of papers in mildly alkaline environment. Polym Degrad Stab 2012;97:118e23. iron gall ink and its risk assessment. Stud Conserv 2009;54:236e54. [31] Ohlsson A, Rydin S. Washing of pulps, part 2. The sorption of Na, Mg and Ca on [6] Strlic M, Radovic T, Kolar J, Pihlar B. Anti- and prooxidative properties of gallic kraft pulp. Svensk Papper 1975;78:549e53. acid in Fenton-type systems. J Agric Food Chem 2002;50:6313e7. [32] ISO 483. Plastics e small enclosures for conditioning and testing using [7] Rouchon V, Duranton M, Burgaud C, Pellizzi E, Lavedrine B, Janssens K, et al. aqueous solutions to maintain the humidity at a constant value 2005. Room-temperature study of iron gall ink impregnated paper degradation [33] TAPPI T 441. Test Cobb: water absorptiveness of sized (non-bibulous) paper under various oxygen and humidity conditions: time-dependent monitoring and paperboard (Cobb test) 2009. by viscosity and X-ray absorption near-edge spectrometry measurements. [34] ISO 5630-3. Paper and board e accelerated ageing e part 3: moist heat Anal Chem 2011;83:2589e97. treatment at 80 C and 65% relative humidity 1996. [8] Burgaud C, Rouchon V, Refait P, Wattiaux A. Mossbauer spectrometry applied [35] TAPPI T 231. Zero-span breaking strength of pulp (dry zero-span tensile) to the study of laboratory samples made of iron gall ink. Appl Phys A Mater Sci 2007. Process 2008;92:257e62. [36] Jerosch H, Lavedrine B, Cherton J-C. Study on the correlation between SEC and [9] Burgaud C, Rouchon V, Wattiaux A, Bleton J, Sabot R, Refait P. Determination mechanical tests of different paper types for degradation state evaluation. of the Fe(II)/Fe(III) ratio in iron gall inks by potentiometry: a preliminary Restaurator 2002;23:222e38. study. J Electroanal Chem 2010;650:16e23. [37] NF Q 03-005. Essais des papiers, cartons et pâtes, détermination du pH des [10] Neevel JG. Phytate: a potential conservation agent for the treatment of ink extraits aqueux 1980. corrosion caused by iron gall inks. Restaurator 1995;16:143e60. [38] ISO 12830. Paper, board and pulps e determination of acid-soluble magne- [11] Rouchon V, Pellizzi E, Duranton M, Vanmeert F, Janssens K. Combining XANES, sium, calcium, manganese, iron, copper, sodium and potassium 2011. ICP-AES and SEM/EDS for the study of phytate chelating treatments used on [39] Stephens CH, Barret T, Whitmore PM, Wade JA, Mazurek J, Schilling M. iron gall ink damages manuscripts. J Anal At Spectrom 2011;26:2434e41. Composition and condition of naturally aged papers. J Am Inst Conserv [12] Reissland B. Ink corrosion aqueous and non aqueous treatment of paper ob- 2008;47:201e15. jects e state of the art. Restaurator 1999;20:167e80. [40] Hill AG, Bishop E, Coles LE, McLauchlan EJ, Meedle W, Pater MJ, et al. General [13] Hahn O, Wilke M, Wolff T. Influence aqueous calcium phytate/calcium method for the determination of iron with 4,7-diphenyl-1,10-phenanthroline hydrogen carbonate treatment on the chemical composition of iron gall inks. (bathophenanthroline). Analyst 1978;103:521e4. Restaurator 2008;29:235e50. [41] Clark LJ. Iron(II) determination in the presence of iron (III) using 4,7-diphenyl- [14] Rouchon V, Burgaud C, Nguyen TP, Eveno M, Pichon L, Salomon J. Iron gall ink 1,10-phenanthroline. Anal Chem 1962;34:348. aqueous treatments: measurement of elemental changes by proton induced [42] Neevel JG, Reissland B. Bathophenanthroline indicator paper. PapierRestaur X-ray emission. PapierRestaur 2008;9:18e28. 2005;6:28e36. [15] Rouchon V, Desroches M, Duplat V, Letouzey M, Stordiau-Pallot J. Methods of [43] Neevel JG. Application issues of the bathophenanthroline test for iron(II) ions. aqueous treatments: the last resort for badly damaged iron gall ink manu- Restaurator 2009;30:3e15. scripts. J Paper Conserv 2012;13:7e13. [44] Jacobi E, Reissland B, Phan Tan Luu C, Van Velzen B, Ligterink F. Rendering the [16] Bray WC, Livingston RS. The catalytic decomposition of hydrogen peroxide in invisible visible. J Paper Conserv 2011;12:25e34. a bromine-bromide solution, and a study of the steady state. J Am Chem Soc [45] Banik G, Bruckle I. Paper and water. Hong Kong: Elsevier; 2011. 1923;45:1251e71. [46] Shahani CJ, Harrison G. Spontaneous formation of acids in the natural aging of [17] Mamou A, Rabani J, Behar D. Oxidation of aqueous Br- by OH radicals studied paper. In: Daniels V, Donnithorne A, Smith P, editors. Works of art on paper by pulse-radiolysis. J Phys Chem 1977;81:1447e8. books, documents and photographs: techniques and conservation. Baltimore: [18] Can ZS, Gurol MD, Iwa. Effect of bromide ions on formaldehyde formation International Institute for Conservation; 2002. p. 189e92. during ozonation of drinking waters. London: I W A Publishing; 2001. [47] Neevel JG. The behaviour of iron and sulphuric acid during iron-gall ink [19] De Laat J, Le TG. Effects of chloride ions on the iron(III)-catalyzed decompo- corrosion. ICOM-CC meeting. Lyon 1999528e33. sition of hydrogen peroxide and on the efficiency of the Fenton-like oxidation [48] Remazeilles C, Quillet V, Calligaro T, Dran JC, Pichon L, Salomon J. PIXE process. Appl Catal B Environ 2006;66:137e46. elemental mapping on original manuscripts with an external microbeam. [20] Bogaard J, Morris HR, Whitmore PM. A method for the aqueous deacidification Application to manuscripts damaged by iron-gall ink corrosion. Nucl Instrum of oxidized paper. J Am Inst Conserv 2005;44:63e74. Methods Phys Res Sect B Beam Interact Mater At 2001;181:681e7.